This document studies the feasibility of utilising Single Antenna Interference Cancellation (SAIC) as a means of
increasing the downlink spectral efficiency of GSM networks.
SAIC is a generic name for techniques, which attempt to cancel or suppress interference by means of signal processing
without the use of multiple antennas. The primary application is the downlink, where terminal space and aesthetics
typically preclude the use of multiple antennas.
Clause 1 of this document defines the scope and objectives of this feasibility study. Clause 4 defines the network
scenarios that have been defined to evaluate SAIC performance in GSM networks. These scenarios are representative
of typical GSM deployments worldwide today. Clause 5 presents the interference statistics associated with the network
scenarios defined in Clause 4. These interference statistics are developed via system simulations, and are defined in
terms of the distributions of the parameters which are critical to understanding SAIC performance. These critical
The Carrier to Interference plus noise Ratio (CIR)
The Dominant to rest of Interferer Ratio (DIR)
The other interferer ratios, which define the relative power of the dominant co-channel interferer to each of the
other considered interferers
The delay between the desired signal and each of the interferers.
It is important to understand the network statistics of these key parameters since most SAIC algorithms can only cancel
one interferer, and their effectiveness in doing this is affected by the 'remaining' interference, and delays between the
desired signal and the interferers.
In Clause 6, candidate SAIC algorithms are evaluated at the link level based on the interference statistics defined in
Clause 5. Both 'long-term average' and per burst results are generated. The long-term average results represent the
classical way of looking at link performance via link simulations, defining the Bit Error Rate (BER) and Frame Error
Rate (FER) averaged over the entire simulation run as a function of the CIR. This is the type of performance that is
typically specified in the GSM standards. However, to develop a system capacity estimate, it is necessary to define the
link performance on a per burst basis. To this end, Clause 6 also defines the average BER over the burst as a function
of the burst CIR and burst DIR. This burst performance is used to develop a link-to-system level mapping. This
mapping is used in Clause 7 to develop voice capacity and data throughput estimates for both conventional and SAIC
receivers. The voice capacity gain and data throughput gain for SAIC is then deduced from these estimates.
Clause 8 describes the field trials that have been conducted using an SAIC prototype Mobile Station (MS). Clause 9
addresses testing considerations for SAIC capable MSs, while Clause 10 defines a couple of signalling options for
identifying an MS as being SAIC capable. Finally, Clause 11 provides the relevant conclusions that can be drawn from
this feasibility study, the most important of which is the conclusion that SAIC is a viable and feasible technology, which
will support significant voice capacity gains for both synchronous and asynchronous networks when applied to GMSK
modulation. In addition, modest increases in GPRS data throughput are also supported for the types of data traffic
considered. Clause 11 also identifies those clauses of the core and testing specifications that will be impacted by the
inclusion of an SAIC capability.
full Table of Contents for TR 45.903 Word version: 15.0.0